Title :
Computational intelligence and cerebellar enigmas
Author :
Dunin-Barkowski, W.L. ; Shishkin, S.L. ; Wunsch, Donald C.
Author_Institution :
Appl. Comput. Intelligence Lab., Texas Tech. Univ., Lubbock, TX
Abstract :
A three-layer cerebellar cortex presents a kind of ideal network for early perceptron theories, but its functioning is much more complicated. Three related topics connected with information coding and memorizing mechanisms in the cerebellum are considered: 1) demonstration that activity transformation by granule cells, conserves the signal continuity provided by snake-in-box type codes; 2) determining precision requirements for the standard CMAC as well as for more realistic cerebellar models to meet demands of adaptive critic based solutions for a nonlinear control problem (swinging pendulum); and 3) demonstration that the closed loop scheme for the synaptic weight adjustment provides a precise setting of cerebellar input-output relations and multiple predictions for physiological experiments
Keywords :
cerebellar model arithmetic computers; encoding; learning (artificial intelligence); neurocontrollers; neurophysiology; tracking; CMAC; adaptive critic design; cerebellar cortex; cerebellar models; closed loop systems; granule cells; information coding; memorizing mechanisms; nonlinear control; perceptron; synaptic weight adjustment; Adaptive control; Arithmetic; Brain modeling; Cameras; Code standards; Computational intelligence; Multilayer perceptrons; Optical fiber testing; Predictive models; Programmable control;
Conference_Titel :
Neural Networks Proceedings, 1998. IEEE World Congress on Computational Intelligence. The 1998 IEEE International Joint Conference on
Conference_Location :
Anchorage, AK
Print_ISBN :
0-7803-4859-1
DOI :
10.1109/IJCNN.1998.682354